[rotrex]

The limit is that you run pure methanol at some point, a fuel that can make a lot of power in a turbo application.
I usually measure the AFR with a wideband lambda gauge. You inject your methanol or mix at the rate you want to try and adjust fuel to get to the AFR you want.
I also suggest to get away from AFRs of different fuels. Use lambda or just to for gasoline AFR you need. Thing is that lambda is a more linear scale with lambda 1.1 running 10% less fuel than lambda 1.0. There or no different lambda values for different fuels. The amounts you needs calculate from the ratio of the nerdy densities. In case of methanol about half of that of gasoline. Ethanol has a energy density of roughly 70% of gasoline.
Adding water with identical combustible fuel floe does not change the lambda value.
Substitution does. But again, just run a wideband lambda sensor and fox fueling on the fly.

OTherwise, you are right. For every 1 flow unit of methanol you add, you need to about remove half a unit of flow of gasoline to maintain the same lambda value or gasoline equivalent AFR in the engine.

Recently I had the impression I can extract more power from my engine with more alcohol and less water. Boost is fixed in my car as it is supercharged.,
Going from 50/50 to 75/25 I was able to advance ignition timing and gained feelable power. Was only a quick drive, but still. I can only assume that not all of the water I inject does actually contribute to knock suppression. The water's action is usually indicted though it's flame speed retardation requiring more ignition timing to compensate.
Now, I could also go in the other direction and reduce methanol and inject more water and less methanol......

It seems like those little direct port jets do not produce the fineest mist. Riceracing's air assisted nozzle in the airbox is definitly a better source of very fine droplets helping water being more efficient in the cylinder. The big droplets plains do not evaporate fast enough in the cylinder. The alcohol does. In a recent publication I read they did experiments and simulations on direct vs port injection of fuels with various alcohol content. The author compared the in cylinder knock suppression and cooling effect by adjusting the intake air tempersature of the turbo engine until both injection methods produced the same knock limited power.
The result was that direction injection was always superior to port injection having more knock suppression. Her attributes this, backed by evaporation models, that much of the ethanol droplets evaporate hitting the hot intake valve.

Now for my direct port scenario this could mean adding methanol does two things:
It adds high octane fuel
And it potentially leaves smaller, but effective water droplets behind as the methanol evaporates.

In the past I did experiments with pure water and never found it to be as effective as a mix with methanol. It's knock suppression performance was worse. Mapping both pure water and mix to its knock limit, timing is vastly different with water needing more advance, the mix made more power. This fits rather well with the water not actually doing its thing that well in my setup. It could very well be a in cylinder distribution issue. Here riceracing's system is superior to even a direct port system as he generates a very very fine mist that follows the bends and evenly distributes in the cylinder almost like a gas. My direct port system does not do that.
Bigger droplets plainly do not evaporate even during the critical part of combustion while the flame front still runs across the cylinder. Any droplets that are not evaporated by the time all the fuel has been consumed by the flame do not contribute to knock suppression.

As boost is in my setup is rather low and intake temps are not that high too due to charge cooling, I believe my setup rather likes the alcohol than the water.

I need to try a air assisted nozzle some time, but I only have 0.5 to about 1 bar of air pressure available to drive it.